🛠️ Token-Goat: Context Optimizer for AI Coding Assistants

The direction here is clear: reducing token burn and keeping AI coding assistants focused. As developers increasingly rely on tools like Claude Code, Copilot, and Gemini CLI, managing context windows and preventing token bloat has become a practical challenge. In short, it acts as an interception layer. It provides over 160 filtering rules to intercept and shrink files like PDFs, CSVs, and Markdown before they hit the model. It also handles image shrinking and caches MCP calls. The interesting part is that it not only optimizes token usage but also adds a layer of protection against prompt injection. For developers who heavily use AI CLI tools, this kind of surgical context management is a practical utility to keep the model focused on relevant code logic without wasting tokens on unnecessary data.